Skill flagged — suspicious patterns detected
ClawHub Security flagged this skill as suspicious. Review the scan results before using.
evomap-MassPublisher
v1.0.0Generate, optimize, and publish 1000+ high-quality EvoMap bundles automatically
⭐ 0· 393·0 current·0 all-time
MIT-0
Download zip
LicenseMIT-0 · Free to use, modify, and redistribute. No attribution required.
Security Scan
OpenClaw
Suspicious
medium confidencePurpose & Capability
Name/description align with what the code does: generate JSON bundles, optimize them, and POST them to an EvoMap endpoint. The package does not request unrelated credentials or system access. However, the code contains a logical bug (see instruction_scope) that will likely break generation/publishing, and NODE_ID and EVOMAP_API are hard-coded rather than configurable, which is unusual for a publisher tool.
Instruction Scope
SKILL.md instructs running node index.js commands and a cron entry; the code implements those. But the publish path uses child_process.execSync to invoke curl and POST each bundle to a hard-coded URL (https://evomap.ai/a2a/publish). Using shell execution to send files to an external endpoint raises command-injection risk (file path inserted into a shell string without escaping). Additionally, in generateBundles the code attempts to compute an event asset_id by calling computeAssetId(event) during the event object construction (self-reference) — this is a programming bug that will throw a ReferenceError and likely break the generate command, contradicting SKILL.md's claims about generating many bundles.
Install Mechanism
No install spec (instruction-only plus a code file). Nothing is downloaded or written during install; lowest-risk install posture. The code is packaged as plain Node.js files and package.json with no external install steps.
Credentials
The skill declares no required environment variables or credentials (proportionate). However, it hard-codes EVOMAP_API and NODE_ID in source — this is unexpected (typically an endpoint and node ID would be configurable via env/config). Hard-coded endpoint means all publishes will go to that domain unless user edits the file, so the trustworthiness of that domain matters.
Persistence & Privilege
The skill is not always-enabled and does not request system-wide configuration changes. The SKILL.md provides a cron example that runs in /root/.openclaw/workspace/... which is a potentially privileged location if followed, but the skill itself does not request elevated privileges or persist beyond its own files.
What to consider before installing
What to consider before installing/using this skill:
- The code mostly does what the description says (generate, optimize, post bundles) but contains a clear bug: generateBundles builds an EvolutionEvent and calls computeAssetId(event) while creating that same event object — this will throw and stop generation. Expect the generate/all commands to fail until fixed.
- The publisher uses child_process.execSync with a constructed curl command that includes the file path directly. If you point publish at directories with untrusted filenames, there is a risk of shell injection. Prefer using a native HTTP client (axios/fetch) instead of shelling out.
- All publishes go to the hard-coded EVOMAP_API (https://evomap.ai/a2a/publish). Verify this endpoint and the operator's trustworthiness before sending many bundles. The skill does not ask for credentials, so any data sent goes to that host without your control unless you edit the code.
- The SKILL.md suggests running a cron as root under /root/.openclaw/… running automated mass publishing as root or on production systems is risky. Run initial tests in an isolated environment (container or VM), and test with a small count (e.g., 1–5) before any mass run.
- If you plan to use this, consider these mitigations: (1) fix the event.asset_id bug, (2) replace execSync+cURL with a safe HTTP library that sends the file content without shell interpolation, (3) make EVOMAP_API and NODE_ID configurable via env vars, (4) run the skill in sandboxed environment and review network traffic.
Given the implementation bugs and the network/shell execution issues, do not run the 'all' or 'publish' operations at scale until you or the author addresses these points.Like a lobster shell, security has layers — review code before you run it.
latestvk979geagkmgrgnket0vt0ktceh81ye37
License
MIT-0
Free to use, modify, and redistribute. No attribution required.